IOP Conference Series: Earth and Environmental Science

PAPER • OPEN ACCESS Related content

- Kinematics analysis of a 3-DOF hydraulic
Kinematics Analysis and Workspace Calculation of driven parallel mechanism
Bing Zhang, Jiandong Li, Ziliang Jiang et
a 3-DOF Manipulator al.

- Robot Kinematics Identification: KUKA

LWR4+ Redundant Manipulator Example
To cite this article: Jianqiang Zhu and Fang Tian 2018 IOP Conf. Ser.: Earth Environ. Sci. 170 Sergey Kolyubin, Leonid Paramonov and
042166 Anton Shiriaev

- Probabilistic scan mode of a robot

manipulator workspace using EEG signals.
Part I
Fernando A Auat Cheeín, Fernando di
View the article online for updates and enhancements. Sciascio, Teodiano Freire Bastos et al.

This content was downloaded from IP address 14.139.241.217 on 20/10/2019 at 20:11

2nd International Symposium on Resource Exploration and Environmental Science IOP Publishing
IOP Conf. Series: Earth and Environmental Science 170 (2018)
1234567890 ‘’“” 042166 doi:10.1088/1755-1315/170/4/042166

Kinematics Analysis and Workspace Calculation of a 3-DOF

Manipulator

Jianqiang Zhu1, a and Fang Tian1, b

1
Shenyang University of Technology, Shenyang 110870, China
a
axios1225@gmail.com, bsytf2003@126.com

Abstract. Robot working space is an important kinematic indicator. Exact computation

of the boundary shape and volume or area of the manipulator workspace is very
important for its optimum design and application. This paper describes the kinematics
analysis of a 3-DOF manipulator and the calculation of its workspace. The kinematics
model of the three degree of freedom (3-DOF) robot was set up with the D-H
parameters method. The effectiveness of kinematics equations was verified by the
Robotics Toolbox simulation of the MATLAB. The robotics workspace of the
manipulator was analyzed and simulated based on the Monte-Carlo method and
MATLAB software programming. The analysis results show that the manipulator’s
working space is well-knit and this method is feasible and practical, prove the
trajectory planning of workspace, and can well perform task, which lay a foundation
to further study of the trajectory planning and motion control.

1. Introduction
The kinematics analysis on the manipulator is an important part of the robot study [1]. The
manipulator is composed of a serial of rigid links connected to each other with revolute or prismatic
joints. Each joint location is usually defined relative to neighboring joint. Calculating the position and
orientation of the end-effector of the manipulator by the given joint angles is known as forward
kinematics [2, 3]. Forward kinematics problem is straightforward and there is no complexity in
deriving the equations. Inverse kinematics is a mapping form the operational space to the joint space,
so it can be considered as the inverse of forward kinematics. An efficient Inverse kinematic algorithm
for robot manipulator is very useful [4, 5]. Working space of robotic arm is the reference point at the
end of the robot arm to achieve the set point of space. Working space is discussed from the geometric
aspects of the work performance of the robot arm. Analytical work is used to determine the spatial
configuration of the robot arm. Working space represents the range of activities, which is a parameter
of the ability to work the robot arm kinematics, and is also an important indicator [6]. Robot arm and
the sizes of work space according to the design requirements, must reach the required working space,
which is reasonable to judge about the structure, operability and flexibility of the robot arms. Work
space model for robot planning and control information provide important constraints, such as
working space was determined according to the mechanical interference between the robot arm and
the robot arm motion planning to avoid obstacles and so on [7, 8].
In the literature, several methods have been proposed to determine the workspace of manipulators
by using analytic, geometric, or numerical approaches. Early studies on the serial manipulator

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
2nd International Symposium on Resource Exploration and Environmental Science IOP Publishing
IOP Conf. Series: Earth and Environmental Science 170 (2018)
1234567890 ‘’“” 042166 doi:10.1088/1755-1315/170/4/042166

workspace were developed by Roth (1975) [9]. A geometric algorithm to compute the parallel
manipulator workspace has been used (Boneveand and Ryu, 2001) [10]. A discretization method is
presented to determine the workspace of both serial and parallel manipulators by Gianni (2008) [11]
and a new software tool designed to compute and allow visualization of the different types of
workspaces of parallel manipulators by E. Macho (2011) [12].
In this paper, the kinematics of a 3-DOF manipulator was studied and proposed a numerical
method to solve the problem. Monte Carlo method [13] was then used to simulate the working space,
which can straightforwardly and visually describe the working space of robot with high speed. It was
indicated that the working points in the workspace of the robot were distributed compactly and
uniformly, which can meet design requirements with high efficiency.

2. Structure of the 3-DOF Manipulator

The manipulator has three degrees of freedom: base, shoulder, elbow, and end-effector. The length of
the Link1 is 1000mm, Link 2 is 550mm, and the distance from center of the end-effector to elbow
(Link3) is 550mm. The structure of the manipulator is described through the Robotics Toolbox Link()
functions [14], in Fig. 1.

Figure 1. The three dimention image of manipulator.

2.1. D-H Coordination System

Denavit and Hartenberg had put forward a matrix method to build the attached coordinate system on
each link in the joint chains of the robot for describing the relationship of translation or rotation
between the contiguous links in 1955. The configuration of the manipulator was built with three
degrees of freedom: base, shoulder, elbow, and wrist, as shown in Fig.2. D-H coordinate system of the
manipulator is found in Fig. 3. Table 1 is the link parameters.

2
2nd International Symposium on Resource Exploration and Environmental Science IOP Publishing
IOP Conf. Series: Earth and Environmental Science 170 (2018)
1234567890 ‘’“” 042166 doi:10.1088/1755-1315/170/4/042166

Figure 2. The structure of the manipulator.

Figure 3. Robot arm coordinate frame of link.

Table 1. Link parameters.

i ai-1(°) αi-1(mm) di(mm) θi(°) Variable range
1 0 0 0 θ1 -180°-180°
2 0 90° 0 θ2 0°-90°
3 a2 0 0 θ3 -60°-60°
4 0 90° d4 0 0

The transformation matrix of the manipulator is given according to Danevit-Hartenberg method.

3
2nd International Symposium on Resource Exploration and Environmental Science IOP Publishing
IOP Conf. Series: Earth and Environmental Science 170 (2018)
1234567890 ‘’“” 042166 doi:10.1088/1755-1315/170/4/042166

轾
c1 - s1 0 0 轾
c2 - s 2 0 0 轾
c3 - s 3 0 a2 轾
1 0 0 0
犏 犏 犏 犏
犏
s1 c1 0 0 犏0 0 - 1 0 犏
s 3 c3 0 0 犏
0 0 - 1 - d4
0
1
T = 犏
犏
1
2
T = 犏
犏
2
3
T = 犏
犏
3
4
T = 犏
犏
0
犏 0 1 0 s
犏2 c 2
0 0 犏0 0 1 0 0
犏 1 0 0
犏 犏 犏 犏
犏
0
臌
0 0 1 犏
臌
0 0 0 1 犏
臌
0 0 0 1 犏
0
臌
0 0 1

3. Forward Kinematic Analysis

The forward kinematic analysis means that the location and pose of the end of manipulator in a given
references coordinates of the links and the variables of the joints for the manipulator. Its kinematic
equation could be expressed as the follows.

轾
nx ox ax px
犏
犏
ny oy ay py
0
4
T = 01T 12T 23T 34T = 犏
犏 (1)
nz
犏 oz az pz
犏
犏
臌
0 0 0 1

where (n x , n y , n z ) is the normal vector, (ax , ay , az ) is a approach vector, (ox , oy , oz ) is a orientation

vector, ( , , ) is the end-effector position vector. Eventually, is
轾
c1c23 s1 c1s 23 c1s 23d 4 + c1c2a2
犏
犏
s1c23 - c1 s1s 23 s1s 23d 4 + s1c2a 2
0
4
T = 犏
犏 (2)
犏s 23 0 - c23 - c23d 4 + s2a2
犏
犏
臌
0 0 0 1

And,
px = c1s 23d 4 + c1c2a 2
py = s1s 23d 4 + s1c2a2
pz = - c23d 4 + s 2a2
When the structural parameter of the manipulator is obtained, the workspace of the manipulator
depends on generalized coordinate . For the limit of the actual structure and driving device, so,
qimin ＃qi qimax . The assemble of point P can be defined as the workspace of the manipulator, namely

ìï x (q )
ïï p i
ï
wR = í y p (qi ) qimin ＃qi qimax (3)
ïï
ïï z p (qi )
ïî

The position of the end-effector of the manipulator relative to the base coordinate system was
obtained from (2).

ìï x (q ) = c s d + c c a
ïï 1 23 4 1 2 2
ï
P = í y (q ) = s1s 23d 4 + s1c2a 2 (4)
ïï
ïï z (q ) = - c23d 4 + s 2a2
îï

Where s23 = sin (2 + 3 ), c23 = cos (2 + 3 ) .

This manipulator is composed of three links and four revolute joints. Most manipulator has limits
on the link length and the range of rotational angle, for this manipulator, the joint limits is:

4
2nd International Symposium on Resource Exploration and Environmental Science IOP Publishing
IOP Conf. Series: Earth and Environmental Science 170 (2018)
1234567890 ‘’“” 042166 doi:10.1088/1755-1315/170/4/042166

- pi ＃1 pi
pi
0 ＃2
2
pi pi
＃ 3
2 3

4. Monte Carlo Method for Constrained Workspace

The Monte Carlo Method for the workspace of manipulator does not use inverse kinematics but needs
only forward kinematics. Constraints including the joint limits can be easily coped by this method. It is
relatively simple and flexible. It is also convenient to use this method to directly obtain graphical
depiction of the workspace [15].

4.1. Simulation
In this paper, the Monte Carlo method was used to determine the workspace of the manipulator.
Values could be yielded randomly in their rotational range, from (5).

imin i = imin + imax - imin ? R and ()

( ) (5)

In, imin : Minimum range of motion joint i, imax : Maximum range of motion joint i, i means joint
number.
These random points from the graphical form, that is, the robot arm workspace cloud, namely the
robot arm of Monte Carlo work space.
Computer programming can be realized by above algorithm, and the results that come out with the
graphics, as shown in Fig 4. Where, n means iteration number.

Figure 4. Working space simulation flow based on Monte Carlo method.

5
2nd International Symposium on Resource Exploration and Environmental Science IOP Publishing
IOP Conf. Series: Earth and Environmental Science 170 (2018)
1234567890 ‘’“” 042166 doi:10.1088/1755-1315/170/4/042166

4.2. Application
The following figures are results of the robot arm work space simulation. Fig 5a describes the end of
the three degrees of freedom robot arm based on the coordinate reference point formed in the
workspace. Fig 5b, Fig 5c and Fig 5d show the three Cartesian coordinates in the base coordinates of
the projection plane, such as XOY, XOZ, YOZ plane.

Figure 5. Workspace.

As the Monte Carlo method itself limits the work space by the robot arm simulation results with the
theoretical value gap. In order to make simulation results more close to the actual working space,
increasing the iteration numbers. We can get as close to the actual working space as possible.

5. Inverse Kinematic Analyses

Inverse kinematics is used to calculate the corresponding joint angles when the posture of the end
effect is known. In this work, the algebraic method is adopted, and we can separate all joint angles.
Namely,

轾
nx ox ax px
犏
轾
0
n 0
o 0
a 0
p 犏
ny oy ay py
0
4
T = 犏
4
犏
4 4 4
= 犏
犏 = 01T (q1 ) 12T (q2 ) 23T (q3 ) 34T (6)
犏0 0 0 0 nz
犏 oz az pz
臌 犏
犏
臌
0 0 0 1

(1) 1

0 - 1
1
T (q1) 04T = 12T (q2 ) 23T (q3 ) 34T (7)

6
2nd International Symposium on Resource Exploration and Environmental Science IOP Publishing
IOP Conf. Series: Earth and Environmental Science 170 (2018)
1234567890 ‘’“” 042166 doi:10.1088/1755-1315/170/4/042166

轾 c1 s1 0 0 轾
nx ox ax px
犏 犏
犏
-犏s1 c1 0 0 犏
ny oy ay py
犏 = 12T (q2 ) 23T (q3 ) 34T = 14T (8)
犏0 0 1 0 犏
nz oz az pz
犏 犏
犏 犏
犏
臌
0 0 0 1 犏
臌
0 0 0 1

In matrix equation (8), element (2, 4) is corresponding equal, so:

- s1px + c1py = 0 (9)

According to the Trigonometric substitution formula, as following:

px = r 1 cos f ; py = r 1 sin f (10)

Where r 1 = px 2 + py 2 , f = arctan( py / px ) ,

ü
sin(f - q1 ) = 0 ïïï
cos(f - q1 ) = ? 1 ïï
ïý (11)
f - q1 = 0 ïï
ïï
q1 = arct an( py / px ) ïïþ

(2) 2
Matrix equation (8) element (1, 4) and (3, 4) is corresponding equal,

c1px + s1py = s 23d 4 + a2c2 ü

ïï
ý (12)
pz = - c23d4 + s2a2 ïï
ïþ

As following,

(
- s2 pz + c2 c1px + s1py = h ) (13)

In, h=
(c1px + s1py ) + pz2 + a 22 - d 42
,
2a 2

- pz = r 2 cos f 2 ; c1px + s1py = r 2 sin f 2 (14)

Where, r 2 = pz 2 + (c1px + s1py )2 ; f 2 = arctan((c1px + s1py ) / (- pz )) ;

sin(f 2 - q2 ) = h / r 2 ü
ïï
ïï
cos(f 2 - q2 ) = ? 1 (h / r 2 ) 2 ïï
ï (15)
ý
f 2 - q2 = arct an((h / r 2 ) / ? 1 (h / r 2 )2 ) ïï
ïï
q2 = arct an((c1px + s1py ) / (- pz ))- arct an(h / ? r 22 h 2 )ïïï
þ

7
2nd International Symposium on Resource Exploration and Environmental Science IOP Publishing
IOP Conf. Series: Earth and Environmental Science 170 (2018)
1234567890 ‘’“” 042166 doi:10.1088/1755-1315/170/4/042166

(3)  3
According to equation (12), the  3 equation showing,

tan(q2 + q3 ) = (c1px + s1py - c2a2 ) / (- pz + s2a2 ) (16)

q3 = arctan((c1px + s1py - c2a2 ) / (- pz + s2a2 )) - q2 (17)

q3 = arct an((c1px + s1py - c2a2 ) / (-pz + s 2a 2 ))

(18)
- arct an((c1px + s1py ) / (-pz )) + arct an(h / ? r 22 - h 2 )

All joint variables are as follows,

q1 = arct an( py / px ) ü
ïï
ï
2 2 ïï
q2 = arct an((c1px + s1py ) / (- pz ))- arct an(h / ? r 2 h ) ïï
ý (19)
q3 = arct an((c1px + s1py - c2a 2 ) / (-pz + s 2a 2 )) ïï
ïï
- arct an((c1px + s1py ) / (-pz )) + arct an(h / ? r 22 - h 2 )ïï
ïþ

According to the workspace that already given on above analysis, it verified the end-effector
location (800, 0, 0) in Robotics Toolbox, as shown in Fig.6.

Figure 6. The end-effector of 3-DOF manipulator at point.

6. Conclusion
In this paper, the forward and inverse kinematics solutions of the manipulator were analyzed. The
range of joint variables was obtained by the simulation of forward kinematics, it lays the foundation
for the workspace analysis, the motion control and the trajectory planning of manipulator in future.
Monte Carlo method was used to simulate the working space, which can straightforwardly and
visually describe the working space of robot with high speed. It was indicated that the working points
in the workspace of the robot were distributed compactly and uniformly, which can meet design
requirements with high efficiency. The effectiveness of the working space model was testified by the
simulation analysis on MATLAB and experimental validation.

8
2nd International Symposium on Resource Exploration and Environmental Science IOP Publishing
IOP Conf. Series: Earth and Environmental Science 170 (2018)
1234567890 ‘’“” 042166 doi:10.1088/1755-1315/170/4/042166

The kinematics of manipulator was simulated in the Robotics toolbox on the workspace, and the
inverse kinematics of manipulator was verified to be correct by the use of a point, and this end-effector
can work well with the manipulator. So the 3-DOF manipulator can meet the working requirement in
factory.

References
[1] Smith‐Guerin N, Nouaille L, Vieyres P, et al. A medical robot kinematics design approach
based on knowledge management [J]. Industrial Robot, 2008, 35(4):316-323.
[2] Chen X, Qin S. Kinematic Modeling for a Class of Free-floating Space Robots J]. IEEE Access,
2017, PP(99):1-1.
[3] He C, Wang S, Xing Y, et al. Kinematics analysis of the coupled tendon-driven robot based on
the product-of-exponentials formula [J]. Mechanism & Machine Theory, 2013, 60(1):90-111.
[4] González-Palacios M A. The unified orthogonal architecture of industrial serial manipulators[J].
Robotics & Computer Integrated Manufacturing, 2013, 29(1):257-271.
[5] Liu H, Zhou W, Lai X, et al. An Efficient Inverse Kinematic Algorithm for a PUMA560-
Structured Robot Manipulator [J]. International Journal of Advanced Robotic Systems, 2013,
10:1.
[6] Yoshida E, Murata S, Kamimura A, et al. A Self-Reconfigurable Modular Robot:
Reconfiguration Planning and Experiments [J]. International Journal of Robotics Research,
2002, 21(10):903-915.
[7] Ciobanu R I, Dobre C, Cristea V. Kinematic and Workspace Analysis of a Parallel Robot Used
in Security Applications [C]. International Conference on Mechatronics, Electronics and
Automotive Engineering, 2014:3-8.
[8] Tao Y, Chen F, Xiong H. Kinematics and Workspace of a 4-DOF Hybrid Palletizing Robot [J].
Advances in Mechanical Engineering, 2014, (2014-6-23), 2014, 2014(8):125973-125973.
[9] Roth B. Performance evaluation of manipulators from a kinematic view point [C], 1970.
[10] Bonev I A, Ryu J. A geometrical method for computing the constant-orientation workspace of
6- ja:math RRS parallel manipulators [J]. Mechanism & Machine Theory, 2001, 36(1):1-13.
[11] Ottaviano E. A Fairly General Algorithm to Evaluate Workspace Characteristics of Serial and
Parallel Manipulators [J]. Mechanics Based Design of Structures & Machines, 2008,
36(1):14-33.
[12] Macho E, Pinto C, Amezua E, et al. Software Tool to Compute, Analyze and Visualize
Workspaces of Parallel Kinematics Robots [J]. Advanced Robotics, 2011, 25(6-7):675-698.
[13] Rubinstein R Y. Simulation and the Monte Carlo Method [M]. John Wiley & Sons, 2008:167-
168.
[14] Corke P. Robotics, Vision and Control [J]. Springer Tracts in Advanced Robotics, 2011, 39(6).
[15] Guan Y, Yokoi K. Reachable Space Generation of A Humanoid Robot Using The Monte Carlo
Method [C]. Ieee/rsj International Conference on Intelligent Robots and Systems,
2006:1984-1989.